set() adds new cols when rows aren't updated

NEWS.md

@@ -178,6 +178,8 @@ d1[d2, on="id", verbose=TRUE]
 
 This feature resolves [#4387](https://github.com/Rdatatable/data.table/issues/4387), [#2947](https://github.com/Rdatatable/data.table/issues/2947), [#4380](https://github.com/Rdatatable/data.table/issues/4380), and [#1321](https://github.com/Rdatatable/data.table/issues/1321). Thanks to @jangorecki, @jan-glx, and @MichaelChirico for the reports and @jangorecki for implementing.
 
+23. `set()` now adds new columns even if no rows are updated, [#5409](https://github.com/Rdatatable/data.table/issues/5409). This behavior is now consistent with `:=`, thanks to @mb706 for the report and @joshhwuu for the fix.
+
 ## TRANSLATIONS
 
 1. Fix a typo in a Mandarin translation of an error message that was hiding the actual error message, [#6172](https://github.com/Rdatatable/data.table/issues/6172). Thanks @trafficfan for the report and @MichaelChirico for the fix.

inst/tests/tests.Rraw

@@ -19002,3 +19002,12 @@ test(2277.2, DT[, closure(b), env=list(closure=var)], 0.5)
 test(2277.3, DT[, closure(b), env=list(closure=stats::var)], 0.5)
 test(2277.4, DT[, closure(b), env=list(closure=stats:::var)], 0.5)
 test(2277.5, DT[, lambda(b), env=list(lambda=function(x) sum(x))], 7L)
+
+# test that set() correctly adds new columns even if no rows are updated
+dt = data.table(a=1L)
+test(2278.1, set(copy(dt), 0L,          "b", logical(0)), data.table(a=1L, b=NA))
+test(2278.2, set(copy(dt), NA_integer_, "b", NA),         data.table(a=1L, b=NA))
+test(2278.3, set(copy(dt), 0L,          "b", NA),         copy(dt)[0L, b := NA])
+test(2278.4, set(copy(dt), NA_integer_, "b", logical(0)), copy(dt)[NA_integer_, b := logical(0)])
+test(2278.5, set(copy(dt), integer(0),  "b", numeric(0)), copy(dt)[integer(0),  b := numeric(0)])
+test(2278.6, { set(dt, 0L, "b", logical(0)); set(dt, 1L, "a", 2L); dt }, data.table(a=2L, b=NA))

man/assign.Rd

@@ -35,7 +35,7 @@ set(x, i = NULL, j, value)
 \item{LHS}{ A character vector of column names (or numeric positions) or a variable that evaluates as such. If the column doesn't exist, it is added, \emph{by reference}. }
 \item{RHS}{ A list of replacement values. It is recycled in the usual way to fill the number of rows satisfying \code{i}, if any.  To remove a column use \code{NULL}. }
 \item{x}{ A \code{data.table}. Or, \code{set()} accepts \code{data.frame}, too. }
-\item{i}{ Optional. Indicates the rows on which the values must be updated with. If not provided, implies \emph{all rows}. The \code{:=} form is more powerful as it allows \emph{subsets} and \code{joins} based add/update columns by reference. See \code{Details}.
+\item{i}{ Optional. Indicates the rows on which the values must be updated. If not \code{NULL}, implies \emph{all rows}. Missing or zero values are ignored. The \code{:=} form is more powerful as it allows adding/updating columns by reference based on \emph{subsets} and \code{joins}. See \code{Details}.
 
     In \code{set}, only integer type is allowed in \code{i} indicating which rows \code{value} should be assigned to. \code{NULL} represents all rows more efficiently than creating a vector such as \code{1:nrow(x)}. }
 \item{j}{ Column name(s) (character) or number(s) (integer) to be assigned \code{value} when column(s) already exist, and only column name(s) if they are to be created. }

src/assign.c

@@ -378,12 +378,13 @@ SEXP assign(SEXP dt, SEXP rows, SEXP cols, SEXP newcolnames, SEXP values)
     for (int i=0; i<targetlen; ++i) {
       if ((rowsd[i]<0 && rowsd[i]!=NA_INTEGER) || rowsd[i]>nrow)
         error(_("i[%d] is %d which is out of range [1,nrow=%d]"), i+1, rowsd[i], nrow);  // set() reaches here (test 2005.2); := reaches the same error in subset.c first
-      if (rowsd[i]>=1) numToDo++;
+      if (rowsd[i]>=0) numToDo++;
     }
     if (verbose) Rprintf(_("Assigning to %d row subset of %d rows\n"), numToDo, nrow);
     // TODO: include in message if any rows are assigned several times (e.g. by=.EACHI with dups in i)
     if (numToDo==0) {
-      if (!length(newcolnames)) {
+      // isString(cols) is exclusive to calls from set()
+      if (!length(newcolnames) && !isString(cols)) {
         *_Last_updated = 0;
         UNPROTECT(protecti);
         return(dt); // all items of rows either 0 or NA. !length(newcolnames) for #759